Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206

 
CONTINUE READING
Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206
Low-Head Hydropower Assessment
                  of the Brazilian State of São Paulo

Open-File Report 2014–1206                              Tietê River
                                                        near Ibitinga,
                                                        São Paulo,
U.S. Department of the Interior                         Brazil
U.S. Geological Survey
Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206
Cover. Portion of a Landsat 8 image showing Tietê River near Ibitinga, São Paulo, Brazil. Image acquired September 2, 2014.
Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206
Low-Head Hydropower Assessment of the
Brazilian State of São Paulo

By Guleid A. Artan, William M. Cushing, Melissa L. Mathis, and Larry L. Tieszen

Open-File Report 2014–1206

U.S. Department of the Interior
U.S. Geological Survey
Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206
U.S. Department of the Interior
SALLY JEWELL, Secretary

U.S. Geological Survey
Suzette M. Kimball, Acting Director

U.S. Geological Survey, Reston, Virginia: 2014

For more information on the USGS—the Federal source for science about the Earth, its natural and living
resources, natural hazards, and the environment, visit http://www.usgs.gov or call 1–888–ASK–USGS.
For an overview of USGS information products, including maps, imagery, and publications,
visit http://www.usgs.gov/pubprod
To order this and other USGS information products, visit http://store.usgs.gov

Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the
U.S. Government.

Although this information product, for the most part, is in the public domain, it also may contain copyrighted materials
as noted in the text. Permission to reproduce copyrighted items must be secured from the copyright owner.

Suggested citation:
Artan, G.A., Cushing, W.M., Mathis, M.L., and Tieszen, L.L., 2014, Low-head hydropower assessment
of the Brazilian State of São Paulo: U.S. Geological Survey Open-File Report 2014–1206, 15 p.,
http://dx.doi.org/10.3133/ofr20141206.

ISSN 2331-1258 (online)
Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206
iii

Contents
Abstract............................................................................................................................................................1
Introduction.....................................................................................................................................................1
     Assessment Area...................................................................................................................................1
     Digital Elevation Model.........................................................................................................................2
     River Network Datasets........................................................................................................................2
     Regionalized Stream Discharge Characteristics.............................................................................2
     Streamflow and Rainfall Data..............................................................................................................2
DEM Data Conditioning..................................................................................................................................3
     Stream Conditioning..............................................................................................................................3
     Stream Segmentation and Drop Estimations....................................................................................5
     Synthetic Streams Contributing Area Compared to Agência Nacional de Águas
                Streamgages Contributing Area............................................................................................5
Assessment of the Streamflow Estimates..................................................................................................5
     Mean Annual Flow.................................................................................................................................6
     Flow Duration Curves............................................................................................................................6
Hydropower Assessment Results..............................................................................................................10
     Hydropower Potential by Basin.........................................................................................................11
     Segment Hydropower Potential........................................................................................................11
     Available Hydropower Potential........................................................................................................12
     Developed and Reaming Hydropower.............................................................................................14
Web Mapping Service And Contour Level Extraction............................................................................14
Summary........................................................................................................................................................15
References Cited..........................................................................................................................................15

Figures
         1.     Map showing the extent of the study area of the State of São Paulo, shown
                with the map of Brazil...................................................................................................................2
         2.     Map showing the hydrologic units within the State of São Paulo........................................3
         3.     Maps showing locations of the Agência Nacional de Águas gages that were
                used to estimate the distribution of rainfall and to validate estimated streamflow...........4
         4.     Maps showing stream networks from the Agência Nacional de Águas dataset
                overlaid on synthetic streams created from the SRTMDEM..................................................5
         5.     Scatterplot with the 1:1 line of upstream area recorded in the Agência Nacional
                de Águas database and contributing area calculated from the digital elevation
                model for the same locations, and bar chart of percent disagreement between
                the two areas..................................................................................................................................6
         6.     Maps showing part of the study area showing mean annual precipitation fields,
                and mean annual precipitation for the same areas when weighted by
                upstream area................................................................................................................................7
         7.     Graphs showing evaluation of the annual average streamflows estimated with the
                hydrologic regionalization model compared to observed long-term annual mean
                streamflows....................................................................................................................................7
Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206
iv

        8.   Graphs showing comparison of Flow Duration Curves (FDCs) calculated from
             log-Pearson distribution fit of observed annual mean flows with FDCs estimated
             from the regionalized hydrologic equation of watersheds in the State of São Paulo.......8
        9.   Graphs showing observed flows at gages with long-term data at 40 percent,
             60 percent, and 70 percent exceedance probability plotted with flows estimated
             with the regionalized hydrologic equations and estimated with the modified
             regionalized equations..................................................................................................................9
       10.   Map showing major streams in São Paulo where the hydropower assessment
             was completed.............................................................................................................................10
       11.   Graph showing three classes of low head/low power technologies.................................11
       12.   Pie diagram showing power category distribution for the potential annual mean
             hydropower of all the streams that are entirely in the State of São Paulo........................12
       13.   Map showing total annual mean hydropower potential of the 21 basins within the
             State of São Paulo.......................................................................................................................12
       14.   Graph showing total low head/low power potential summarized by basin.......................13
       15.   Map showing distribution of available power potential (annual mean power) of
             São Paulo energy resources among three hydropower classes........................................13
       16.   The user interface of the Web mapping site that was developed for extraction
             of contour lines from the DEM, and an example of an extracted contour line
             product..........................................................................................................................................14

     Table
        1.   Hydropower classes...................................................................................................................11
Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206
v

Conversion Factors
SI to Inch/Pound
               Multiply                         By                             To obtain
                                              Length
 millimeter (mm)                              0.03937             inch (in.)
 meter (m)                                    3.281               foot (ft)
 kilometer (km)                               0.6214              mile (mi)
                                               Area
 square meter (m2)                            0.0002471           acre
 square kilometer (km )
                      2
                                           247.1                  acre
 square meter (m2)                           10.76                square foot (ft2)
 square kilometer (km )
                      2
                                              0.3861              square mile (mi2)
                                             Flow rate
 cubic meter per second (m3/s)               35.31                cubic foot per second (ft3/s)
 cubic meter per second per square           91.49                cubic foot per second per square
   kilometer [(m3/s)/km2]                                           mile [(ft3/s)/mi2]

Elevation, as used in this report, refers to distance above the vertical datum.
Vertical and horizontal coordinate information is referenced to the World Geodetic System
1984 (WGS 84).
NOTE TO USGS USERS: Use of hectare (ha) as an alternative name for square hectometer
(hm2) is restricted to the measurement of small land or water areas. Use of liter (L) as a special
name for cubic decimeter (dm3) is restricted to the measurement of liquids and gases. No prefix
other than milli should be used with liter. Metric ton (t) as a name for megagram (Mg) should be
restricted to commercial usage, and no prefixes should be used with it.
Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206
Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206
Low-Head Hydropower Assessment of the Brazilian State
of São Paulo

By Guleid A. Artan,1 William M. Cushing,2 Melissa L. Mathis,3 and Larry L. Tieszen2

Abstract                                                                    already been harnessed. In addition to the annual mean hydro-
                                                                            power estimates, potential hydropower estimates with flow
      This study produced a comprehensive estimate of the                   rates with exceedance probabilities of 40 percent, 60 percent,
magnitude of hydropower potential available in the streams                  and 90 percent were made.
that drain watersheds entirely within the State of São Paulo,
Brazil. Because a large part of the contributing area is out-
side of São Paulo, the main stem of the Paraná River was                    Introduction
excluded from the assessment. Potential head drops were
calculated from the Digital Terrain Elevation Data,which has                      This report describes methods and processes used to
a 1-arc-second resolution (approximately 30-meter resolu-                   produce a hydropower assessment for the Brazilian State of
tion at the equator). For the conditioning and validation of                São Paulo by the U.S. Geological Survey (USGS). The work
synthetic stream channels derived from the Digital Elevation                was done on behalf of the Corporación Andina de Fomento
Model datasets, hydrography data (in digital format) sup-                   (CAF) – Development Bank of Latin America. The assessment
plied by the São Paulo State Department of Energy and the                   emphasizes low-head hydropower potential. Typically turbines
Agência Nacional de Águas were used. Within the study area                  that operate on streams with drops in elevation that are less
there were 1,424 rain gages and 123 streamgages with long-                  than 20 meters (m) are classified as low-head hydropower.
term data records. To estimate average yearly streamflow, a                 Low-head hydropower plants usually do not require damming
hydrologic regionalization system that divides the State into               of the streams; therefore, they have fewer negative environ-
21 homogeneous basins was used. Stream segments, upstream                   mental impacts than high-head hydropower plants and they
areas, and mean annual rainfall were estimated using geo-                   offer the opportunity to provide off-grid power in remote areas
graphic information systems techniques. The accuracy of the                 where the installation of power grids often is prohibitively
flows estimated with the regionalization models was validated.              expensive. Assessment of hydropower potential requires cal-
Overall, simulated streamflows were significantly correlated                culating the drops in elevation and estimating the potentially
with the observed flows but with a consistent underestimation               available streamflow with associated exceedance probability
bias. When the annual mean flows from the regionalization                   levels for all stream segments that are within the study area.
models were adjusted upward by 10 percent, average stream-                        In the subsequent sections of this report, we introduce
flow estimation bias was reduced from -13 percent to -4 per-                the data and methods used to produce a hydropower poten-
cent. The sum of all the validated stream reach mean annual                 tial assessment for the Brazilian State of São Paulo. The
hydropower potentials in the 21 basins is 7,000 megawatts                   objective of the work was to build datasets that can be used
(MW). Hydropower potential is mainly concentrated near                      for detailed feasibility studies on the development of small
the Serra do Mar mountain range and along the Tietê River.                  hydropower plants; these studies will be done by a Brazilian
The power potential along the Tietê River is mainly at sites                consulting firm.
with medium and high potentials, sites where hydropower has

                                                                            Assessment Area
  1
    InuTeq ASRC, Contractor to the U.S. Geological Survey; work performed        The hydropower assessment was completed for all the
under contract G13PC00028.
                                                                            streams located entirely within the State of São Paulo (fig. 1).
  2
   U.S. Geological Survey.                                                  We excluded streams that the State of São Paulo shares with
  3
   SGT, Inc., Contractor to the U.S. Geological Survey; work performed      other Brazilian states; for example, no assessment was carried
under contract G10PC00044.                                                  out for the main stem of the Paraná River.
Low-Head Hydropower Assessment of the Brazilian State of São Paulo - Open-File Report 2014-1206
2   Low-Head Hydropower Assessment of the Brazilian State of São Paulo

Figure 1. The extent of the study area
(shaded black) is the State of São Paulo,
shown with the map of Brazil.

                                                                                                                           At
                                                                                                                           lan
                                                                                                                            tic
                                                                                                                                Oce
                                                                                   BRAZIL

                                                                                                                                 an
                                                      Pa
                                                      cif
                                                        ic
                                                           Ocean
                                                                                                   São Paulo

                                                  N                                               Paraná River
                                                                                                       0       500     1,000 KILOMETERS

                                                                                                       0             500         1,000 MILES

Digital Elevation Model                                            Regionalized Stream Discharge Characteristics
      Potential head drops and stream connectivity were calcu-           Streamflow data used to assess hydropower potential
lated using a Digital Elevation Model (DEM) from the Shuttle       were produced using the hydrologic regionalization method
Radar Topography Mission (SRTM) Digital Terrain Elevation          of São Paulo State (Liazi and others, 1988). The hydrologic
Data (DTED2), which has a 1-arc-second resolution (approxi-        regionalization divides the State into 21 homogeneous regions
mately 30-m at the equator) (Farr and others, 2007). The data      (fig. 2). Estimating flow rates using the São Paulo State
                                                                   regionalization method on the synthetic streams required two
are from the National Aeronautics and Space Administration
                                                                   inputs: the spatial extent and the mean annual rainfall of the
(NASA) and the National Geospatial-Intelligence Agency
                                                                   upstream contributing area to the location of interest. The
(NGA), and are archived by the USGS Earth Resources
                                                                   hydrologic regionalization models for estimating the mean
Observation and Science (EROS) Center. The USGS and NGA            annual flow rate (Liazi and others, 1988) are of the form
have an agreement to improve the accuracy of the DEM and
to archive the data. Even though the SRTM DEM is restricted                            Q=(a+b*P)*Area/1,000                               (1)
from public access, this agreement allows the USGS to pro-
cess the full resolution 1-arc-second DEM data and to create       where
derivative products for most of South and Central Americas.                 Q      is the mean annual flow (m3/sec),
                                                                       a and b     are model parameters that are basin specific,
                                                                             P     is the mean annual precipitation of the
River Network Datasets                                                                 contributing watershed area above the
                                                                                       segment end in millimeters (mm), and
      The São Paulo State Department of Energy and the                     Area    is the contributing area to the stream segment
Agência Nacional de Águas (ANA) supplied the USGS with                                 where the computation is done (km2).
hydrography data (in digital format). This data layer was used
for the conditioning and validation of synthetic stream chan-
nels derived from the DEM datasets. The ANA hydrography            Streamflow and Rainfall Data
data layer was digitized from 411 topographic sheets at a
                                                                         Stream and rain gage data used for this study are from
1:50,000 scale in a horizontal datum South American Datum
                                                                   the ANA hydrological information system Website Sistema
1969 (SAD-69) and Lambert Conformal Conic projection.              de Informações Hidrológicas-HidroWeb (http://hidroweb.ana.
To have more confidence the synthetic streams derived from         gov.br). Data used were from 1,424 rain gages with at least
the DEM, which is the basis for the hydropower assess-             10 years of data records and 123 streamgages with at least
ment, the ANA hydrography layer was used to validate the           28 years of data records. The rain gage stations were well
drainage network of the synthetic streams and to eliminate         distributed throughout the study area and sufficient to char-
erroneous streams.                                                 acterize annual rainfall (fig. 3A). The streamgages were more
DEM Data Conditioning   3

densely distributed in the southeastern section and had mini-   In cases where the synthetic streams deviated substantially
mal coverage in western portions of São Paulo (see fig. 3B).    from the ANA hydrography layer, a conditioning process was
     The spatial mean annual rainfall for the study area was    applied to better align synthetic streams to the topographic
estimated from the ANA rainfall database using an inverse       hydrographic data. Figure 4 shows an overlay of synthetic
distance weighted interpolation method. The most recent         streams from the DEM and the hydrography dataset from
streamgage data are from the early 1990s.                       ANA in part of the study area. In general, there is good spatial
                                                                agreement between the two datasets, excluding the fact that
                                                                the synthetic stream reaches were shorter than the ANA
                                                                hydrography reaches in the upstream areas. The shortness of
DEM Data Conditioning                                           the synthetic streams was a result of setting the stream cells’
                                                                generation threshold higher to avoid the inclusion of “false”
      Potential head drops and stream connectivity were
                                                                synthetic stream reaches. False synthetic stream reaches would
calculated from the DTED2 DEM dataset. For some areas,
                                                                have inflated the power potential estimates.
especially areas with low relief, synthetic streams derived
                                                                      Five DEM conditioning steps were made: (1) areas in the
from DEM data could diverge from the real conditions hydro-
                                                                DEM that needed modifications were identified by compar-
graphically present on the ground. Hydrological conditioning
                                                                ing synthetic streams with the ANA-supplied hydrography
is the process of modifying elevation pixels from the DEM to
                                                                data, (2) the stream segments that intersect those areas were
improve stream channel modeling that more closely represents
                                                                selected, (3) DEM pixels underlying areas of the selected
ground conditions.
                                                                stream segments were set to zero, (4) synthetic stream net-
                                                                works were created from the DEM with the modified pixels,
Stream Conditioning                                             and (5) using the synthetic stream networks from the modified
                                                                DEM and the original DEM, all elevation pixels underly-
     For areas having moderate to high relief and a well-       ing the network were reduced by 20 m and then hydrologic
developed drainage network, the watershed data derived from     fill processing was applied. This relative depth conditioning
the DEM generally worked well. In areas with low relief, the    technique allows the enforcement of natural stream chan-
synthetic hydrography derived from the first DEM process-       nels with minimal alteration of relative elevation measure-
ing deviated from the ANA-supplied hydrographic network.        ments. Because all pixels intersecting the stream network

                                                                                                                           an
                                                                                                                     Oce
                                                                                                                 c
                                                                                                        la   nti
                                                                                                   At
                                                                               HENRY BORDEN DAM

   N
                                                                                       0     100         200 KILOMETERS

                                                                                       0           100                200 MILES

Figure 2. The hydrologic units within the State of São Paulo.
4   Low-Head Hydropower Assessment of the Brazilian State of São Paulo

       A
                                                                                                                        EXPLANATION
                                                                                                                          Rain gage network

                                                                                                                          ean
                                                                                                                     Oc
                                                                                                                ic
                                                                                                           nt
                                                                                                    la
                                                                                                 At
        N
                                                                                                  0              100          200 KILOMETERS

                                                                                                  0                     100            200 MILES

       B
                                                                                                                       EXPLANATION
                                                                                                                       Streamgage network

                                                                                                                          ean
                                                                                                                     Oc
                                                                                                                ic
                                                                                                           nt
                                                                                                      la
                                                                                                 At
        N
                                                                                                  0              100          200 KILOMETERS

                                                                                                  0                     100            200 MILES

     Figure 3. Locations of the Agência Nacional de Águas gages that were used to estimate the distribution of rainfall and to
     validate estimated streamflow: A, rain gages, and B, streamgages.
Assessment of the Streamflow Estimates   5

                 EXPLANATION
                                                              N
                   Synthetic stream

                   Cartographic stream

                                 0    100 200 KILOMETERS                      0      2       4 KILOMETERS

                                 0       100   200 MILES                      0          2             4 MILES

          Figure 4. Stream networks from the Agência Nacional de Águas dataset overlaid on synthetic streams created
          from the SRTMDEM.

were modified, their elevation relative to one another does not    Synthetic Streams Contributing Area Compared
change, which allows for consistent elevation drop calcula-        to Agência Nacional de Águas Streamgages
tions on the stream network.
                                                                   Contributing Area
Stream Segmentation and Drop Estimations                                 The synthetic stream network created from the DEM
                                                                   was visually compared with ANA-generated hydrographic
      Elevation drop is a fundamental component in estimat-        data and quantitatively compared with watershed areas in the
ing low-head hydropower potential. The drop is calculated at       ANA streamgage database. There was good spatial agreement
intervals of 1-km segments along the synthetic streams pre-        between upstream areas recorded in the ANA database and
sented in the Stream Conditioning section. Stream segmenting       those estimated from the DEM data. The comparison indicated
starts at the mouth or at a confluence of other streams with the   that there were no large discrepancies in drainage area from
                                                                   the two sources (fig. 5A); however, although the agreement
stream of interest; some segments can be less than 1 km, but
                                                                   was within 2 percent on average, there were a few small basins
this only happens at the end of a stream or downstream from a
                                                                   where the discrepancy between the two area estimates was
confluence. This process continues upstream in every synthetic
                                                                   greater than 20 percent (fig. 5B).
stream until the points have been equally distributed along the
stream. During this segmentation process, a unique identifi-
cation is given to every stream segment, which is used as a
key for all the attributes of the segments. The upstream and       Assessment of the Streamflow
downstream points of the stream segments created in the first      Estimates
step are used to extract the upstream and downstream eleva-
tion values from the DEM. Elevation drops were calculated by            Streamflows were estimated with hydrologic regionaliza-
taking the difference between the upstream and downstream          tion models. Estimated streamflow was validated by compari-
elevation measurements of the segments. Calculated elevation       som with average flows calculated from gages with long-term
drops assigned to the stream segments are part of the hydro-       data records. The flow validation was extended to the stream-
power assessment database attribute table.                         flows estimated from the developed Flow Duration Curves.
6   Low-Head Hydropower Assessment of the Brazilian State of São Paulo

    Mean Annual Flow                                                                                                                      simulated streamflows were significantly correlated with the
                                                                                                                                          observed flows but with a consistent underestimation bias
          The equations for the hydrologic regionalization of Liazi                                                                       (fig. 7A). When the annual mean flows from the regionaliza-
    and others (1988), described in the “Streamflow and Rainfall                                                                          tion models (Liazi and others, 1988) were adjusted upward by
    Data” section, were applied at every stream segment. Stream                                                                           10 percent, average streamflow estimation bias reduced from
    segments, upstream areas, and mean annual rainfall were                                                                               -13 percent to -4 percent (fig. 7B).
    estimated using geographic information systems (GIS) tech-
    niques. Pixel-to-pixel flow direction and flow accumulation
    calculated from the 30-m DEM provided the framework for
                                                                                                                                          Flow Duration Curves
    the calculation of the upstream mean rainfall for every pixel                                                                                In addition to the average yearly streamflows used to
    on the stream segments.                                                                                                               estimate annual potential hydropower, several values of the
          The continuous parameterization technique described                                                                             streamflow Flow Duration Curves (FDCs) were calculated.
    by Harvey and Eash (1996) was applied to calculate the                                                                                The FDCs are usually shown as a plot of a percentage of time
    upstream area-weighted mean annual rainfall of every pixel                                                                            that streamflow is likely to equal or exceed a specified value
    on the streams. This parameterization technique uses the same                                                                         of interest. With the given FDC dataset, users should be able
    rationale as the flow accumulation algorithm, but instead of                                                                          to calculate reliability levels for various hydropower potential
    treating every pixel as one unit, each pixel is weighted by the                                                                       levels. For example, the FDC can be used to show the percent-
    rainfall value calculated for the pixel with inverse distance                                                                         age of time river flows can be expected to exceed a design
    weighted interpolation. Figures 6A–B show the differences                                                                             flow of some specified value, or to show the discharge of the
    between the local mean annual rainfall and area-weighted                                                                              stream that occurs or is exceeded some percentage of the time
    mean annual rainfall values for a neighborhood of several                                                                             (for example, 90 percent of the time).
    pixels in the study area.                                                                                                                    The FDCs of the segments were predicted using equa-
          The accuracy of the flows estimated with the regional-                                                                          tions from the São Paulo regionalization method (Liazi and
    ization method for São Paulo was validated by comparing                                                                               others, 1988). Predicted FDCs were validated by comparing
    the estimated flows to the ANA streamgage-observed flows                                                                              those estimated with a log-Pearson Type III distribution fit to
    from sites with long-term data records (more than 28 years).                                                                          ANA streamgage datasets. The log-Pearson distribution was
    The ANA streamgage database contains 119 gages, of which                                                                              fit to only those streamgages with at least 28-year records. For
    34 gages had long-term data records. Average annual flows                                                                             the hydropower assessment, the shapes of the FDCs in low-
    and flows with exceedance probabilities of 40 percent, 60 per-                                                                        flow regions are of most interest. The shapes of the curves in
    cent, and 90 percent estimated with the regionalization method                                                                        the high-flow regions indicate the type of likely flood regime,
    were compared to observed streamflow with corresponding                                                                               whereas the shapes of the low-flow regions characterize the
    exceedance probabilities. Figure 7A shows the comparison                                                                              ability of the basin to sustain low-flow regimes during dry
    of the mean streamflow estimates from the regionalization                                                                             seasons. The median flow is the discharge that is exceeded
    equations with the flows from the ANA database. Overall,                                                                              50 percent of the time. The basic time unit of flow used in

                                                               A                                                                                                                B
                                                      80,000                                                                                                              30
Digital elevation model (DEM), in square kilometers

                                                                   R2 = 0.9997
                                                                                                                                                                          20
                                                      60,000
                                                                                                                                            Area difference, in percent

                                                                                                                                                                          10

                                                      40,000                                                                                                               0

                                                                                                                                                                          -10
                                                      20,000
                                                                                                                                                                          -20

                                                          0                                                                                                               -30
                                                               0                 20,000       40,000           60,000            80,000                                         0   5   10   15      20       25   30   35   40

                                                                    Agência Nacional de Águas (ANA) area, in square kilometers                                                                    Gage site

    Figure 5. Scatterplot with the 1:1 line of upstream area recorded in the Agência Nacional de Águas database and contributing area
    calculated from the digital elevation model for the same locations; and B, bar chart of percent disagreement between the two areas.
Assessment of the Streamflow Estimates   7

preparing the FDC affects the appearance of the FDC. When                                                            Included in the estimated flows database delivered with
mean flows for a 1-year period are used instead of daily val-                                                  the final product is the value of the 7-day minimum flow
ues, as in this case, the averaging flattens the resulting FDC.                                                expected in the stream segments. The estimates of 7-day
      Figures 8A–O depict the FDC estimates from the region-                                                   minimum flows could be used to estimate flows that should
alization method at 15 selected sites from the ANA streamflow                                                  be left in the streams for ecological protection and subtracted
                                                                                                               from the discharge that goes into the calculation of the hydro-
database. The FDCs from the observed streamflow data were
                                                                                                               power. We did not validate the 7-day minimum flow estimates
from the annual average flows fitted to a log-Pearson distribu-
                                                                                                               because of a lack of observed minimum flow data, but from
tion. Overall, the two FDC estimates agree for most parts of
                                                                                                               a literature review (Rezende and others, 2010) we concluded
the curves, but there is constant underestimated bias in the                                                   that the minimum flow estimates seem to have poor predic-
FDCs estimated from the regionalization method for the parts                                                   tive skills. The most recent studies of the observed streamflow
of the FDC that represent the low-flow regime conditions. The                                                  in the ANA database were from the early 1990s; hence, there
regionalized hydrologic equations underestimation bias of the                                                  could be new water abstraction developments that are not
low-flow regimes is more apparent in figures 9A–F.                                                             reflected in either of the observed or estimated flow data.

                                                  A                                                              B

                Figure 6. Part of the study area showing A, mean annual precipitation fields, and B, mean annual precipitation for the same
                areas when weighted by upstream area.

                                                          A                                                           B
                                                  400                                                           400
                                                              R2 = 0.98                                                   R2 = 0.98
    Simulated flows, in cubic meters per second

                                                              Bias = -13 percent                                          Bias = -4 percent
                                                  300                                                           300

                                                  200                                                           200

                                                  100                                                           100

                                                      0                                                           0
                                                          0              100       200   300             400          0              100        200         300         400

                                                                                           Observed flows, in cubic meters per second

    Figure 7. Evaluation of the annual average streamflows estimated with the hydrologic regionalization model compared to
    observed long-term annual mean streamflows: A, streamflows estimated with original equations, and B, streamflows estimated
    with improved regression equations derived from available stream and rain gage data of the study area.
8   Low-Head Hydropower Assessment of the Brazilian State of São Paulo

                                         30

                                         25                                                                                                             EXPLANATION
                                         20                                                                                                                  Observed

                                         15                                                                                                                  Simulated

                                         10

                                         5
                                                  A. Site 58030000                     B. Site 61895000                              C. Site 62496000
                                         0
                                        100

                                         80

                                         60

                                         40

                                         20

                                                  D. Site 62540000                     E. Site 62600000                              F. Site 62675100
                                         0
Discharge, in cubic meters per second

                                        140

                                        120

                                        100

                                         80

                                         60

                                         40

                                         20
                                                  G. Site 58105000                     H. Site 58110000                              I. Site 62496000
                                          0
                                        180
                                        160
                                        140
                                        120
                                        100
                                         80
                                         60
                                         40
                                         20
                                                  J. Site 61782000                     K. Site 61817000                              L. Site 62410000
                                         0
                                        300

                                        250

                                        200

                                        150

                                        100

                                         50
                                                  M. Site 58158000                     N. Site 61830000                              O. Site 62707000
                                          0
                                              0         20      40   60   80   100 0         20       40      60       80    100 0         20      40   60       80      100
                                                                                            Percent time flow was exceeded

Figure 8. Comparison of Flow Duration Curves (FDCs) calculated from log-Pearson distribution fit of observed annual mean flows with
FDCs estimated from the regionalized hydrologic equation of watersheds in the State of São Paulo (Liazi and others, 1988).
Assessment of the Streamflow Estimates   9

                                                                             Original regressions                                                Improved new regressions

                                                        A. 40-percent exceedance probability                                      B. 40-percent exceedance probability
                                              600

                                              400

                                              200

                                                0
                                                    0                 200                  400                 600            0                 200                 400           600
Simulated flows, in cubic meters per second

                                                        C. 60-percent exceedance probability                                      D. 60-percent exceedance probability
                                              600

                                              400

                                              200

                                                0
                                                    0                 200                  400                 600            0                 200                 400           600

                                              600
                                                        E.   70-percent exceedance probability                                    F.   70-percent exceedance probability
                                              500

                                              400

                                              300

                                              200

                                              100

                                                0
                                                    0          100          200      300         400          500       600   0          100          200     300          400   500    600
                                                                                                       Observed flows, in cubic meters per second

Figure 9. Observed flows at gages with long-term data at 40 percent, 60 percent, and 70 percent exceedance probability
plotted with flows estimated with the regionalized hydrologic equations on the right (A, C, E ), and estimated with the modified
regionalized equations on the left (B, D, F ).
10   Low-Head Hydropower Assessment of the Brazilian State of São Paulo

Hydropower Assessment Results                                                  For potential hydropower, turbine efficiency was assumed
                                                                         to be 100 percent and no hydraulic head loss was considered
      The study produced a comprehensive estimate of the                 in the calculation; therefore, the resulting hydropower esti-
magnitude of hydropower potential available in the streams               mates are for gross potential.
that drain watersheds entirely in the State of São Paulo. Water-               In addition to the annual mean hydropower estimates,
shed assessment locations are shown in figure 10. Streams                potential hydropower estimates with mean annual flows
in São Paulo but with sizeable contributing areas outside of             estimated using an improved set of regression equations, and
the borders of the State of São Paulo were excluded from the             with flow rates with exceedance probabilities of 40 percent,
analysis and are shown in light blue in figure 10. The hydro-            60 percent, and 90 percent, are given with the hydropower
power potential of each stream reach was calculated using the            assessment datasets delivered to CAF.
hydraulic head and annual mean flow rates at the inlet of the                  The results of the hydropower assessment are presented
reach estimated using the regionalization model (Liazi and               with an emphasis on five power classes shown in table 1. The
others, 1988). The hydraulic head associated with each stream            sum of the first four classes is equal to the total power and
reach was obtained using the elevation data in the DEM data-             represents classification by power class, and the low head/
set. The hydropower for the river segments was then estimated            low power class is a classification by power technology (see
as                                                                       fig. 11). The final product of the hydropower assessment is a
                                                                         GIS streams vector layer segmented every 1 km with attributes
                                    P=g*H*Q*10−3                  (2)    containing estimated flows (m3/s), head drop (m), and power
                                                                         potentials (Megawatts (MW)) at the four above-described lev-
where                                                                    els of the probability. Figure 12 presents a summary of results
             P         is hydropower potential [megawatt],               for average power potential of the four power categories for
             H         is hydraulic head [m] of the stream segment,      the streams considered in the analysis.
             Q         is the flow rate of the water in the segment            The accuracy of the power potential estimates is depen-
                           [m3/s], and                                   dent on the accuracy of the individual stream reach power
              g        is gravitational acceleration of 9.81 m/s2.       potentials that were summed to produce total values. The

                                                                                                                              ean
                                                                                                                         Oc
                                                                                                                    ic
                                                                                                               nt
                                                                                                          la                        N
                                                                                                     At

        Base from U.S. National Park Service       0        100         200 KILOMETERS

                                                   0              100                    200 MILES

     Figure 10. Major streams in São Paulo where the hydropower assessment was completed (shown in dark blue).
Hydropower Assessment Results   11

calculated reach flow rates had a standard error of 13 per-                                              Hydropower Potential by Basin
cent. The DEM data, for a random discrete location in South
America, had an absolute height error of 6.2 m (Farr, 2007;                                                    Potential hydropower was aggregated by basin for the
Rodríguez, 2006), but the analyses indicate that the uncer-                                              State of São Paulo (fig. 2). The total power potential for
tainty in the difference between two elevations in near proxim-                                          each of the four basic power classes was calculated by add-
ity (hydraulic head drop) is much better than the elevation                                              ing power potential totals for each power category within
uncertainty for an individual location. Because of the direct                                            each basin. A GIS layer containing the basin boundaries was
relation between power potential and flow rate, the standard                                             intersected with a hydropower assessment dataset of stream
error of the reach power potential would be at least 13 percent.                                         reaches. Figure 13 shows total annual mean power potentials
The uncertainty of the calculated hydraulic head values fur-                                             of the 21 hydrologic regions stacked by power categories. The
ther increases the uncertainty of the power potential values.                                            sum of all the validated stream reach mean annual hydropower
However, if the errors are uniformly distributed, the accuracy                                           potentials in the 21 basins is 7,000 MW. Hydropower potential
of a total value produced by summing a large number of reach                                             is mainly concentrated near the Serra do Mar mountain range
power potentials will be better than the accuracy associated                                             and along the Tietê River (fig. 10). The power potential along
with the individual values that were summed.                                                             the Tietê River is mainly at sites with medium and high poten-
                                                                                                         tials, sites where hydropower has already been harnessed.

                                                                                                         Segment Hydropower Potential
Table 1. Hydropower classes.
                                                                                                              Stream segments with power potentials less than 1 MW
[, greater than]
                                                                                                         and more than 100 kilowatts (kW) with hydraulic heads less
                                                                                                         than 9 m were summed to provide an estimate of total low
                            Hydropower class                       Power class                           head/low power potential at watersheds and regional levels
      Micro                                             < 100 Kilowatt.                                  (Hall and others, 2004). Low head/low power potential is
                                                                                                         presented as separate categories in figure 14 because to exploit
      Mini                                              100 to 1,000 Kilowatt.                           low head/low power potential requires the deployment of
      Small                                             1 to 50 Megawatt.                                unconventional systems or microhydro technology that could
      Medium and large                                  > 50 Megawatt.                                   increase the cost of developing hydropower potential. Basins
                                                                                                         G and P (fig. 2) had the highest potential for low head/low
      Low head/low power                                < 1 Megawatt and < 9.1 meter drop.
                                                                                                         power potential among the 21 basins (fig. 14).

                             14
                                                                                                                         EXPLANATION
                                                                                                                               100 Kilowatt
                             12                                                                                                1 Megawatt

                             10
Hydraulic head, in meters

                              8

                              6             Conventional
                                              turbines

                              4

                              2                                                     Unconventional systems
                                      Microhydro                          (ultra-low head and kinetic energy turbines)
                                                                                                                                                   Figure 11. Three classes
                              0                                                                                                                    of low head/low power
                                  0                10              20                30                40                 50                  60
                                                                                                                                                   technologies (from Hall and
                                                                   Flow rate, in cubic meters per second                                           others, 2004).
12   Low-Head Hydropower Assessment of the Brazilian State of São Paulo

                                                                                                   Available Hydropower Potential
                                                                                                         The amount of potential available hydropower is equal
                                                                                                   to the potential hydropower minus the hydropower already
                                                                                                   developed. For every stream, we calculated developed
                                                                                                   hydropower from the ANA hydroelectric database coverage.
                                                                                                   Figure 15 summarizes the undeveloped hydropower potential
                                                                                                   within the watersheds in the State of São Paulo, subdivided
                                                                                                   into three categories. In the undeveloped streams, there is no
                                                                                                   site within the medium and large power category. A few sites
                                                                                                   along the major rivers (Tietê and do Peixe) have small eleva-
                                                                                                   tion drops and large enough flows to possibly achieve low-
                                                                                                   head hydropower within the medium power categories.
                                                                                                         We determined that the hydropower capacity within the
                                                                                                   study area as provided by the ANA hydroelectric database
                                                                                                   overestimated the amount of developed potential power for
                                                                                                   some sites. Some of the developed plant capacities in the ANA
                                                                                                   hydroelectric database could not be justified on the natural
      Figure 12. Power category distribution for the
                                                                                                   average annual streamflow rates alone. Power capacities of
      potential annual mean hydropower of all the
                                                                                                   some plants were greater than the average power that could be
      streams that are entirely in the State of São Paulo.
                                                                                                   warranted using average flow rates and the hydraulic head of
                                                                                                   the plant’s location (for example, Henry Borden Dam, fig. 2).
                                                                                                         To produce an estimate of developed power potential
                                                                                                   that is comparable to the potential power estimates, which
                                                                                                   are based on annual mean flow rates, it will be necessary to
                                                                                                   estimate the average energy generated by each hydroelectric
                                                                                                   plant, a task that is beyond the scope of this study. A drawback
                                                                                                   of using actual power generated is that efficiency of power
                                                                                                   generation is not included in the numbers estimated with such
                                                                                                   a method.

                                                                                                                                                EXPLANATION
                                                                                 180 (P)
           N                                                                                                             180 (P)   Total annual mean hydropower potential by basin—
                                                                                                                                      Power megawatts (basin code): 180 (P)
                                               39 (U)                                                                              Hydropower category

                                 785 (T)                                                                                           Micro (less than 100 kilowatts)

                                                                                                                                   Mini (100 to 1,000 kilowatts)
                            64 (S)
                                                                    215 (N)        479 (O)                                         Small (1 to 50 megawatts)
               54 (R)
                                                                                                                                   Medium (greater than 50 megawatts)
                                                         526 (M)
                                                                                                                                   Total hydropower: 7,000 megawatts
                        60 (Q)
                                               88 (L)                                          65 (K)
                                                                                                               862 (H)

                                                        392 (J)

                                     123 (I)                             401 (F)              1,278 (B)   98 (C)
                                                                                     86 (E)
                                               584 (D)
                                                                        78 (A)

                                     0            100              200 KILOMETERS

                                     0                      100                    200 MILES

    Figure 13. Total annual mean hydropower potential of the 21 basins within the State of São Paulo.
Hydropower Assessment Results   13

                                     120

                                     100

                                      80
            Low head, in megawatts

                                      60

                                      40

                                      20

                                      0
                                               C        E         A      F       K       B     T        U        Q      O       I        S      R      D      J      H      L      M     N   P       G
                                                                                                                              Basin

           Figure 14. Total low head/low power potential summarized by basin (see fig. 2 for basin locations).

                                                                                                                                                                    EXPLANATION
                                                                                                       133 (P)
       N                                                                                                                                     133 (P)   Available power potential (annual mean power) by basin—
                                                                                                                                                          Power megawatts (basin code): 133 (P)

                                                                      39 (U)                                                                           Hydropower category
                                                                                                                                                       Micro (less than 100 kilowatts)
                                                       785 (T)
                                                                                                                                                       Mini (100 to 1,000 kilowatts)
                                                   64 (S)                                 196 (N)   210 (O)                                            Small (1 to 50 megawatts
           54 (R)
                                                                                                                                                       Total available hydropower: 3,700 megawatts
                                                                                89 (M)
                                           58 (Q)
                                                                      84 (L)                                         60 (K)
                                                                                                                                         633 (H)

                                                            116 (I)                                                374 (B)
                                                                                              86 (E)                                98 (C)
                                                                                                             126 (F)
                                                                      580 (D)
                                                                                               78 (A)

                                                            0            100             200 KILOMETERS

                                                            0                      100                      200 MILES

Figure 15. Distribution of available power potential (annual mean power) of São Paulo energy resources among three
hydropower classes.
14   Low-Head Hydropower Assessment of the Brazilian State of São Paulo

Developed and Reaming Hydropower                                    a stream segment and identify its estimated power potential
                                                                    and related attributes (fig. 16). In addition to presenting the
     To determine the final potential hydropower that can be        information as a Web map, the data also are distributed as
developed, areas where hydropower cannot be developed (for          an Esri Map Service (http://www.geosur.info/map-viewer/
example, parks, wildlife refuges) need to be identified and         hydropower).
excluded. At the time of this report, maps of areas excluded              CAF and the consulting firm IX Estudos e Projetos
from hydropower development in São Paulo were not avail-            needed access to the 1-arc-second (30-m) resolution DEM of
able. If available, an excluded area data layer could be inter-     the region for a follow-up detailed hydropower assessment for
sected with the potential hydropower data layer to identify         the high potential areas identified in this study. The restricted
stream reaches that should be excluded from consideration as
                                                                    nature of the SRTM data presented a unique challenge to the
potential sources of hydropower. In the final product, stream
                                                                    feasibility phase of the project. With permission from NGA,
reaches would be coded as either excluded or not excluded
                                                                    the USGS developed a Web geoprocessing service that pro-
from hydropower development.
                                                                    vides a method for collaborating agencies to extract contours
                                                                    from the DEM. The extraction is done through a Web map-
                                                                    ping site on the Esri Geoprocessing Server (http://tps.geosur.
Web Mapping Service And Contour                                     info/arcgis/rest/services/Models/GeoSUR_HydroDerivatives/
Level Extraction                                                    GPServer). The contour extraction is limited to a 20-km2 area
                                                                    surrounding the selected stream. The contour intervals are set
     In addition to the GIS streams layer, the data are avail-      to 2 m and range between the minimum elevations at the pro-
able in a Web mapping service for collaborating agencies to         spective dam location and a height specified by the user. After
view and extract information. This service provides access          running the Web mapping tool, the contour lines are delivered
to an interactive map of São Paulo that allows users to select      as a downloadable Esri shapefile product (fig. 16).

A                                                                             B

Figure 16. The user interface of the Web mapping site (A) that was developed for extraction of contour lines from the DEM, and an
example of an extracted contour line product (B).
References Cited  15

Summary                                                         Hall, D.G., Cherry, S.J., Reeves, K.S., Lee, R.D., Carroll,
                                                                  G.R., Sommers, G.L., and Verdin, K.L., 2004, Water energy
     A comprehensive assessment of the hydropower potential       resources of the United States with emphasis on low head/
available in all the watersheds in the Brazilian State of São     low power resources, Idaho Falls, Idaho: U.S. Department
Paulo was completed. Potential head drops were calculated         of Energy, Energy Efficiency and Renewable Energy, Wind
from a digital elevation model with a 1-arc-second resolution     and Hydropower Technologies, Idaho National Engineering
(approximately 30-meter resolution at equator). The hydro-
                                                                  and Environmental Laboratory.
meteorological data (rainfall and discharge) used come from
a dense network of gages. Streamflow was estimated using        Harvey, C., and Eash, D., 1996, Description of BASINSOFT,
21 hydrologic regionalization functions. The potential hydro-
                                                                  a computer program to quantify drainage basin charac-
power of all streams was estimated as 7,000 megawatts. The
potential hydropower is mostly concentrated near the Serra do     teristics: Proceedings of the Sixteenth Annual ESRI User
Mar mountain range and along the Tietê River, areas where         Conference,Palm Springs, California, accessed Decem-
most hydropower has already been harnessed. To estimate           ber 2010 at http://proceedings.esri.com/library/userconf/
the potential for low head/low power, stream segments             proc96/to100/pap072/p72.htm.
with power potentials less than 1 megawatt and more than
100 kilowatts with hydraulic heads less than 9 meters were      Liazi, A., Conejo, J.L., Palos, J.C.F., and Cintra, P.S., 1988,
aggregated. Basins G and P showed the highest potential for       Regionalização Hidrológica no estado de São Paulo,
low head/low power.                                               Revista Águas e Energia Elétrica DAEE, v. 5, no. 14,
                                                                  p. 4–10.

References Cited                                                Rezende J.H., Pires, J.S.R., and Mendiondo, E.M., 2010,
                                                                  Hydrologic pulses and remaining natural vegetation in Jaú
Farr, T.G., Rosen, P.A., Caro, E., Crippen, R., Duren, R.,        and Jacaré-Pepira watersheds: Brazilian Archives of Biol-
  Hensley, S., Kobrick, M., Paller, M., Rodriguez, E., Roth,      ogy and Technology, v. 53, no. 5, p. 1127–1136.
  L., Seal, D., Shaffer, S., Shimada, J., Umland, J., Werner,
  M., Oskin, M., Burbank, D., and Alsdorf, D., 2007, The        Rodríguez, E., Morris, C.S, and Belz, J.E., 2006, A global
  Shuttle Radar Topography Mission: Reviews of Geophysics,        assessment of the SRTM performance: Photogrammetric
  v. 45, no. 2, p. 21–22.                                         Engineering & Remote Sensing, v. 72, no. 3, p. 249–260.
Publishing support provided by:
  Rolla Publishing Service Center

For more information concerning this publication, contact:
   U.S. Geological Survey Earth Resources Observation and
   Science (EROS) Center
   47914 252nd Street
   Sioux Falls, South Dakota 57198
   (605) 594-6151

Or visit the EROS Center Web site at:
   http://eros.usgs.gov/
Artan and others—Low-Head Hydropower Assessment of the Brazilian State of São Paulo—Open-File Report 2014–1206

                                                                                                                                           http://dx.doi.org/10.3133/ofr20141206
                                                                                                                 ISSN 2331-1258 (online)
You can also read